logo
Welcome Guest! To enable all features please Login or Register.

Notification

Icon
Error
2 Pages<12
Linda Rascke Holy Grail System - Linda Rascke Holy Grail System
Options
Go to last post Go to first unread
Previous Topic Next Topic
oztrader  
#21 Posted : Tuesday, September 4, 2007 6:12:43 PM(UTC)
oztrader

Rank: Advanced Member

Groups: Registered, Registered Users, Unverified Users
Joined: 7/12/2007(UTC)
Posts: 134
Location: Perth Western Australia

Perhaps consider using a Turnover filter instead of a Volume filter.

Eg. LLV(Mov(V*C,21,E)21)>150000; {$150000 is only a suggestion}.

Regards

oztrader
Back to top
veritech1  
#22 Posted : Wednesday, September 5, 2007 8:49:32 AM(UTC)
veritech1

Rank: Member

Groups: Registered, Registered Users, Subscribers
Joined: 7/9/2006(UTC)
Posts: 14
Cool, however when we start adding to the strategy, we move away from the idea she had. It does not matter if the trend has already started. In fact, thats what we wanted. Thus wanting the ADX(14) to be above 30. We are just waiting for a pullback so that the price will retest low or the high. If the trend is strong enough it will do just that. I noticed however that she does not mention any crossover of any moving averages, the only moving average involved is the 20 day EMA. I think with a ADX less than 20, you will pick stock that are consolidating. This will not be a good idea since there is no sense in direction. We want to at least have an idea where the stock is already heading.
Back to top
johnl  
#23 Posted : Wednesday, September 5, 2007 9:29:30 AM(UTC)
johnl

Rank: Advanced Member

Groups: Registered, Registered Users
Joined: 11/7/2005(UTC)
Posts: 602
Here's my last shot with ADX < 20 before I look at ADX > 30 again:
I added a volume spike (p4,a5) and used a turn in the ADX line to kick in my trade (E series).
I conditioned the signal with Close trending up (e4).
It looks pretty good but you have to watch out for the "Trough" signal since it uses the Zig
function to find the trough. So the assumption is that the ADX doesn't wobble (kick in signals
that dissapear on you later).

p1 := Mov(C,3,S)-Mov(C,10,S);
p2 := ADX(14);
p3 := Mov(C,20,E);
p4 := Mov(V/C,3,E)-Mov(V/C,10,E);

a1:= HHVBars(p1,20);
a2:= BarsSince(Cross(p2,30)=1);
a3:= BarsSince(Cross(p3,L)=1);
a4:= Cross(C,p3);
a5:= HHV(p4,10);

c1:= a4*If((a1<30),1,0)*If((a2>a1),1,0)*
If((a3<a2),1,0)*If((a5>80),1,0);
c4:= If((p2<20),1,0);
d1:=c1*c4;

e1:= BarsSince(d1=1);
e2:= TroughBars(1,p2,10);
e3:= TroughBars(2,p2,10);
e4:= If((C-Ref(C,-3)>0),1,0);

f1:= If((e2<e1),1,0);
f2:= If((e3<e1),0,1);
f3:= If((f1*f2=1) AND (Ref((f1*f2),-1)=0),1,0);
f3*e4
Back to top
johnl  
#24 Posted : Wednesday, September 5, 2007 12:31:04 PM(UTC)
johnl

Rank: Advanced Member

Groups: Registered, Registered Users
Joined: 11/7/2005(UTC)
Posts: 602
Here is someting closer to Linda's requirements:
momentum had to have happened before adx cross over 30
adx has to be above 25
close trend has to be increasing
close has to be above 10 day low and below 25 day high
low has to have dipped below 20 EMA after momentum and adx signal:

p1 := Mov(C,3,S)-Mov(C,10,S);
p2 := ADX(14);
p3 := Mov(C,20,E);
p4 := C-Ref(C,-3);

a1:= BarsSince(Cross(p1,0)=1);
a2:= BarsSince(Cross(p2,30)=1);
a3:= BarsSince(Cross(L,p3)=1);
a4:= If((C>LLV(C,10)) AND (C<HHV(C,25)) AND
(TroughBars(1,C,5)<PeakBars(1,C,5)),1,0);

b1:= If((a1<24),1,0)*
If((a2>a1),1,0)*
If((a3<a2),1,0)*
If((p4>0),1,0)*
If((p2>25),1,0)*a4;

e1:= If((b1=1) AND (Ref(b1,-1)=0),1,0);
e2:= Cross(C,ValueWhen(1,e1=1,C)*1.05);
e2
============================

To test I used a exploration like below to see what happens 6 days after the signal:

Col A: Close CLOSE
Col B: backbb Fml("holy_grail_xx_00")
Col C: R1 (C-Ref(C,-1))*
Ref(Fml("holy_grail_xx_00"),-1)
Col D: R2 (C-Ref(C,-2))*
Ref(Fml("holy_grail_xx_00"),-2)
Col E: R4 (C-Ref(C,-4))*
Ref(Fml("holy_grail_xx_00"),-4)
+
(C-Ref(C,-3))*
Ref(Fml("holy_grail_xx_00"),-3)
Col F: R6 (C-Ref(C,-6))*
Ref(Fml("holy_grail_xx_00"),-6)
+
(C-Ref(C,-5))*
Ref(Fml("holy_grail_xx_00"),-5)
Filter When(HHV(Fml("holy_grail_xx_00"),6)=1)

Filter enabled Yes
Periodicity Daily
Records required 750



Back to top
amory  
#25 Posted : Thursday, September 6, 2007 7:08:17 AM(UTC)
amory

Rank: Advanced Member

Groups: Registered, Registered Users
Joined: 4/18/2007(UTC)
Posts: 99
Location: sydney australia

hi John. I think this last formula of yours is as close as you'll get. it's the kind of exploration where you still have to look at the resulting items & see if they conform to the concept. lucky if you can pick two or three good ones. some explorations are like that, when the idea behind them is very complicated.

what I mean is, you can write a perfectly valid exploration for some winners, e. g. where the SP breaks thru a rising 150MA (Weinstein) on improving volume & the whole formula only takes a couple of lines.

wouldn't you agree, Linda is a lot more demanding,

best regards
Back to top
johnl  
#26 Posted : Thursday, September 6, 2007 7:41:04 AM(UTC)
johnl

Rank: Advanced Member

Groups: Registered, Registered Users
Joined: 11/7/2005(UTC)
Posts: 602
I agree, good entry points on a rising stock, and on a declining stock it usually stays away.
Doesn't get you in at intermediate highs (I hate that) so you should see a profit right off the bat
and makes stop positioning relatively easy. It should get above and stay above that 20 day EMA.
Will use fundementals (growth rate, earning momentum) to direct me to which stocks to buy.
Also looks good if you like to pyramid your trades.
I can see why she likes it.
My testing showed a win/loss ratio a bit above 50/50 which is accepable, just means you still have
to pick the right stock to trade.

So now you all know how to create a new indicator, many attempts, several blind alleys, couple of dead ends, and sometimes something that actually might work.

Sure wouldn't want to try programming this outside the Metastock formula language, would have been a mess.
Back to top
amory  
#27 Posted : Thursday, September 6, 2007 4:46:58 PM(UTC)
amory

Rank: Advanced Member

Groups: Registered, Registered Users
Joined: 4/18/2007(UTC)
Posts: 99
Location: sydney australia

in general terms then, what would the ideal Rascke formation look like on chart? given that there are certain specifics re DI & MA, but other than that are we looking for a Flag & Triangle pattern? because this has proved quite valuable in the past, when it works (not all the time). Linda's way might be designed to remove some of the uncertainty.

anyway, it would be nice to see a typical Rascke chart.

Back to top
veritech1  
#28 Posted : Thursday, September 6, 2007 10:19:02 PM(UTC)
veritech1

Rank: Member

Groups: Registered, Registered Users, Subscribers
Joined: 7/9/2006(UTC)
Posts: 14

I think what happen was the addition of extra indicators made it more complicated. Here is the instructions straight from the book. If you notice there are only six steps. This is her strategy for the Holy Grail. She mentioned that she named it as the Holy Grail only as a joke. However, she called it that because they are easy to come by.

FOR BUYS (SELLS ARE REVERSED)

1. A 14-period ADX must initially be greater than 30 and rising. This will

identify a strongly trending market.

2. Look for a retracement in price to the 20-period exponential moving average.

Usually the price retracement will be accompanied by a turndown in the ADX.

3. When the price touches the 20-period exponential moving average, put a buy

stop above the high of the previous bar.

4. Once filled, enter a protective sell stop at the newly formed swing low. Trail

the stop as profits accrue and look to exit at the most recent swing high. If you

think the market may continue its move, you might exit part of the position at

the most recent swing high and tighten stops on the balance.

5. If stopped out, re-enter this trade by placing a new buy stop at the original

entry price.

6. After a successful trade, the ADX must once again turn up above 30 before

another retracement to the moving average can be traded.
Back to top
amory  
#29 Posted : Friday, September 7, 2007 4:00:45 AM(UTC)
amory

Rank: Advanced Member

Groups: Registered, Registered Users
Joined: 4/18/2007(UTC)
Posts: 99
Location: sydney australia

Thanks Veritech for the definition. but however, not as easy to formulize as it looks. I do think John's is a very good effort towards that goal.

what I posted before about the flag & triangle, that should have read 'flagpole & triangle'. the flagpole being the approaching rise (how else are you going to get that strong ADX) followed by a dip & subsequent up-break out of the triangle (or maybe pennant). at least, that's how I understand its practical application.

Back to top
johnl  
#30 Posted : Friday, September 7, 2007 11:10:05 PM(UTC)
johnl

Rank: Advanced Member

Groups: Registered, Registered Users
Joined: 11/7/2005(UTC)
Posts: 602
I think the last code covers her basics. I included two steps: one being that the adx crossing 30
has to come after a high in momentum (see wabbit's chart reference above) and secondly
the code:
a4:= If((C>LLV(C,10)) AND (C<HHV(C,25)) AND
(TroughBars(1,C,5)<PeakBars(1,C,5)),1,0);
which says the "signal" has to be above the lowest low of the last 10 periods and below the
highest high of the previous 25 periods and the low has to come after the high, in other words
the signal hits the 20 day EMA at end of a 'flagpole & triangle' pattern. The momentum creates
the flagpole and the adx crossing 30 and trailing off creates the triangle.
Without the "a4" code I believe the win/loss drops to around 40% which stinks.
Most of your losses happen not at the beginning but at the end of the trend, would be nice to rid
the indicator of those trades too.

Back to top
amory  
#31 Posted : Saturday, September 8, 2007 4:09:02 AM(UTC)
amory

Rank: Advanced Member

Groups: Registered, Registered Users
Joined: 4/18/2007(UTC)
Posts: 99
Location: sydney australia

hi John. so we are in agreement on the a4:= code & how it aims at a flagpole & triangle, which in itself is a wellknown & proven pattern. plus certain (Linda's) refinements.

but among the results there are also a number of straightout new-highs without the triangle, in which case I would identify them as Ross-hooks. another proven pattern, consisting of one or several higher lows as against more or less equal-level highs, followed by an up-break on decent volume. one would imagine both patterns fall within the specifications of the formula.

by the way, this is how I solve the Volume problem. I use a simple exploration specifying certain basic requirements e.g. >50EMA & a minimum volume or rising volume (or whatever) in conjunction with "Grail" or any other exploration. MS is quite happy to do that, as long as you click "use results of previous explo". it sounds silly, but works for me. saves cluttering-up a good formula.

finally, I don't understand your last sentence, about losses at end of the trend. which trend is it you are referring to? the best formula in the world will not give you better than a three-to-one success ratio. unless the market is in a very positive mood, but I don't think that will apply on monday.

have a nice weekend!
Back to top
veritech1  
#32 Posted : Tuesday, September 11, 2007 5:31:09 PM(UTC)
veritech1

Rank: Member

Groups: Registered, Registered Users, Subscribers
Joined: 7/9/2006(UTC)
Posts: 14
So, when I put the exploration and only get a few outcomes. Are these only buy signals or are they both?
Back to top
veritech1  
#33 Posted : Wednesday, September 12, 2007 3:59:28 PM(UTC)
veritech1

Rank: Member

Groups: Registered, Registered Users, Subscribers
Joined: 7/9/2006(UTC)
Posts: 14
Well I also noticed that even when I ran the exploration. I received only a few stocks, yet, the stock that I chose had an ADX (14) < 30 even though I had changed it from 25 just to see what happened. Not sure why I'm getting this even though it says specifically in the code.
Back to top
johnl  
#34 Posted : Wednesday, September 12, 2007 10:06:43 PM(UTC)
johnl

Rank: Advanced Member

Groups: Registered, Registered Users
Joined: 11/7/2005(UTC)
Posts: 602
I haven't replied to amory yet, need to look at some stuff first.
If you post what you have, I am willing to copy and run to see what results I get.
[Cntl-C copies everything highlighted... Cntl-V pastes everthing highlighted]
Good to see that you are fiddling with the code and looking at the different results.
Back to top
johnl  
#35 Posted : Friday, September 14, 2007 10:13:47 AM(UTC)
johnl

Rank: Advanced Member

Groups: Registered, Registered Users
Joined: 11/7/2005(UTC)
Posts: 602

Here is what I was talking about being able to identify some of the bad trades at the end of the
trend. Open the following chart in another window. The red spikes are the buy signals. The red
line in the middle window is a MA of the spikes, and the blue is a MA of that line. The signals in
May of 2004 are what I am looking at where they kick in just as the trend turns down. The back
line is an ADX.

hxxp://www.pictureuploaderstop.com/user_uploads/johnl745c/AAON.png (change xx to tt)

When the red line crosses the blue (or the blue peaks) is where a lot of the bad trades are so I
was looking to rid me of those trades (increasing my % wins ratio) . Notice that the ADX starts to
move up around that same time which would hint that a new trend is starting. This may be one way
of avoiding a bunch of bad trades..
Back to top
amory  
#36 Posted : Friday, September 14, 2007 9:21:15 PM(UTC)
amory

Rank: Advanced Member

Groups: Registered, Registered Users
Joined: 4/18/2007(UTC)
Posts: 99
Location: sydney australia

John, are you not getting yourself side-tracked a little? a good formula is not easy to construct, especially when it's a complicated one like Linda's & you've done a top job. I would say it's the entry that's important to define hopefully with some precision, exit is a different story. there are any number of indicators that will tell you when the trend is going to change. unless you are planning to use the grail as a mechanical system, but I'm not sure that's advisable.

the results I've been getting on exploration, some of them look promising. altho once the ADX & the DI's come into it, these ones seem to be all over the place. can't really see where they conform to any rigid requirements. sometimes there's a textbook crossover of the DI+ vs the - but the ADX is down low somewhere (e.g. ANN on the ASX, at other times vice versa (e.g. AWE on the ASX)

haven't yet put it to the practical test, becoz of market uncertainty (rates etc).

with best regards ... amory
Back to top
amory  
#37 Posted : Saturday, September 15, 2007 3:33:14 AM(UTC)
amory

Rank: Advanced Member

Groups: Registered, Registered Users
Joined: 4/18/2007(UTC)
Posts: 99
Location: sydney australia

if I may add another observation: I have tried out the "testing" exploration which followed your latest Grail formula. the results I got were identical to those from the Grail exploration itself. is that the way it's meant to come out?

another thing: this "testing" explo contains a number of columns. but only column B is used in the filter. is that intentional?

hope I'm not upsetting you with all these enquiries.
Back to top
veritech1  
#38 Posted : Monday, September 17, 2007 7:20:42 PM(UTC)
veritech1

Rank: Member

Groups: Registered, Registered Users, Subscribers
Joined: 7/9/2006(UTC)
Posts: 14
Interesting. I only use one column. m I using it right?
Back to top
amory  
#39 Posted : Monday, September 17, 2007 9:23:44 PM(UTC)
amory

Rank: Advanced Member

Groups: Registered, Registered Users
Joined: 4/18/2007(UTC)
Posts: 99
Location: sydney australia
hi Veritech. I was merely asking, because John writes a keen formula. I sort of expected some reference to the columns C D E F in the Filter.
Back to top
Users browsing this topic
Guest (Hidden)
2 Pages<12
Forum Jump  
You cannot post new topics in this forum.
You cannot reply to topics in this forum.
You cannot delete your posts in this forum.
You cannot edit your posts in this forum.
You cannot create polls in this forum.
You cannot vote in polls in this forum.

Powered by YAF.NET | YAF.NET © 2003-2025, Yet Another Forum.NET